6 ms·

Almost all of Facebook's "open source" code has the same PATENTS file, which makes them toxic if you can imagine a future where Facebook infringes one of your p

Almost all of Facebook's "open source" code has the same PATENTS file, which makes them toxic if you can imagine a future where Facebook infringes one of your patents: https://github.com/search?p=5&q=org%3Afacebook+filename%3APATENTS&type=Code&utf8=%E2%9C%93 https://github.com/search?p=5&q=org%3Afacebook+filename%3APA... This includes things far outside the React ecosystem: - Flow (JS type checker, like TypeScript) - prepack (JS optimizing transpiler) - a bunch of Android/iOS UI/debugging frameworks, - all their GraphQL libraries (server/client) - their machine learning work (mostly targeting Torch) - Reason (statically typed programming language that transpiles to JS/OCaml) - ZSTD, a highly competitive compression algorithm. Disclaimer: I work at Google on Kubernetes (which is Apache licensed). This is my personal opinion. I am frustrated with software with trivially incompatible licenses.

sanxiyn 9y ago
I am very happy to report that ZSTD removed PATENTS clause: https://github.com/facebook/zstd/pull/801 https://github.com/facebook/zstd/pull/801
- rockdoe 9y ago
  This is huge.
  - markyuckerberg 9y ago
    Welcome to the "folks who miss the forest for the trees" club.
    - coldtea 9y ago
      Welcome to the over-reaction party. 3 years from now, most startups and companies will still be using React as they did for the last 5 years that the same terms held.
      - markyuckerberg 9y ago
        >> Welcome to the over-reaction party Hmmm.. in that case the people who actually need this advice even more would be the maintainers of ZSTD who seem to have hastily changed the license after seeing the backlash from the Apache policy change. Care to stand by this opinion and leave your comment on the thread which discusses the license change over on GitHub? https://github.com/facebook/zstd/issues/775 https://github.com/facebook/zstd/issues/775
- hobofan 9y ago
  RocksDB too. That really makes it seem like Facebooks stance is: "As long as we have a few really huge projects that infiltrate most companies' codebases, our 'smaller' projects can drop the patent clause".
- jnbiche 9y ago
  What the hell? None of this makes any sense. What is Facebook up too?
IanKelling 9y ago
> I am frustrated with software with trivially incompatible licenses. It's not a copyright license, and it's not incompatible with any other free software license. I'm frustrated with the misinformation spread about it. If this patent license/grant get's revoked, you are back to simply using the BSD license with no patent grant. I've read so many people say "you'd have to stop using react if you sued facebook", uh, no, you'd have a bsd license with no patent grant like you probably do with tons of other free software your company uses. Clearly, people should be complaining about that if they are complaining about this, but the misunderstanding and misinformation is really strong. If you believe software patents are universally bad, like many people including me, then it is clearly better using the MIT/BSD license alone, which gives you zero patent rights, you are simply infringing and waiting to be sued. I have no problem with it. https://www.gnu.org/philosophy/software-patents.en.html https://www.gnu.org/philosophy/software-patents.en.html.
- Scaevolus 9y ago
  A patent grant is a license. The patent license is indeed independent of the copyright license, but it's troubling enough that multiple legal teams (Apache and Google, at least) don't want to entangle themselves with it. I'm doubtful that any licensing scheme can significantly drive patent reform. Lobbying and legal precedents hold that capability.
  - IanKelling 9y ago
    > multiple legal teams (Apache and Google, at least) don't want to entangle themselves with it. Citation needed on Google. Apache isn't "not wanting to be entangled", they reject EVERY license which gives stronger protections for user freedoms than apache license alone, like the gpl, lgpl, cddl, mpl, etc, etc, etc. Google has no problem using the gpl for linux and lots of licenses and projects that apache would reject, so you can't just lump them together like it's the same thing.
    - sanxiyn 9y ago
      DannyBee is on this thread. That is good enough citation for me.
diakritikal 9y ago
What's the difference between say the PATENTS file/grant in these Facebook projects and the PATENTS file/grant in Go which Kubernetes is written in? Not being inflammatory, genuinely interested as a user of both Go and React...
- Scaevolus 9y ago
  Not a lawyer, but as I understand it: The Go and Kubernetes patent licenses terminate if you file litigation specifically regarding those projects-- if you sue anyone for violating some garbage collector patent in Go, you lose the patent licenses Go granted. Facebook's patent license terminates if you sue Facebook, subsidiaries, or "any party relating to the Software" for any patent infringement. It doesn't have to be related to React.
  - diakritikal 9y ago
    Thanks, very interesting difference. Google's grant protects (or tries to protect) specific tech, Facebook's grant protects (or tries to protect) as an umbrella, the organisation. Well - can't comment on how effective it will be as a disincentive to sue, but I do feel at the same time if it does succeed in that goal it will almost certainly act equally as a negative factor in adoption; because it's my opinion that this is creating a soft-walled garden around Facebook open source tech.

sanxiyn 9y ago

I am very happy to report that ZSTD removed PATENTS clause: https://github.com/facebook/zstd/pull/801 https://github.com/facebook/zstd/pull/801

rockdoe 9y ago
This is huge.
- markyuckerberg 9y ago
  Welcome to the "folks who miss the forest for the trees" club.
  - coldtea 9y ago
    Welcome to the over-reaction party. 3 years from now, most startups and companies will still be using React as they did for the last 5 years that the same terms held.
    - markyuckerberg 9y ago
      >> Welcome to the over-reaction party Hmmm.. in that case the people who actually need this advice even more would be the maintainers of ZSTD who seem to have hastily changed the license after seeing the backlash from the Apache policy change. Care to stand by this opinion and leave your comment on the thread which discusses the license change over on GitHub? https://github.com/facebook/zstd/issues/775 https://github.com/facebook/zstd/issues/775
hobofan 9y ago
RocksDB too. That really makes it seem like Facebooks stance is: "As long as we have a few really huge projects that infiltrate most companies' codebases, our 'smaller' projects can drop the patent clause".
jnbiche 9y ago
What the hell? None of this makes any sense. What is Facebook up too?

IanKelling 9y ago

> I am frustrated with software with trivially incompatible licenses. It's not a copyright license, and it's not incompatible with any other free software license. I'm frustrated with the misinformation spread about it. If this patent license/grant get's revoked, you are back to simply using the BSD license with no patent grant. I've read so many people say "you'd have to stop using react if you sued facebook", uh, no, you'd have a bsd license with no patent grant like you probably do with tons of other free software your company uses. Clearly, people should be complaining about that if they are complaining about this, but the misunderstanding and misinformation is really strong. If you believe software patents are universally bad, like many people including me, then it is clearly better using the MIT/BSD license alone, which gives you zero patent rights, you are simply infringing and waiting to be sued. I have no problem with it. https://www.gnu.org/philosophy/software-patents.en.html https://www.gnu.org/philosophy/software-patents.en.html.

Scaevolus 9y ago
A patent grant is a license. The patent license is indeed independent of the copyright license, but it's troubling enough that multiple legal teams (Apache and Google, at least) don't want to entangle themselves with it. I'm doubtful that any licensing scheme can significantly drive patent reform. Lobbying and legal precedents hold that capability.
- IanKelling 9y ago
  > multiple legal teams (Apache and Google, at least) don't want to entangle themselves with it. Citation needed on Google. Apache isn't "not wanting to be entangled", they reject EVERY license which gives stronger protections for user freedoms than apache license alone, like the gpl, lgpl, cddl, mpl, etc, etc, etc. Google has no problem using the gpl for linux and lots of licenses and projects that apache would reject, so you can't just lump them together like it's the same thing.
  - sanxiyn 9y ago
    DannyBee is on this thread. That is good enough citation for me.

diakritikal 9y ago

What's the difference between say the PATENTS file/grant in these Facebook projects and the PATENTS file/grant in Go which Kubernetes is written in? Not being inflammatory, genuinely interested as a user of both Go and React...

Scaevolus 9y ago
Not a lawyer, but as I understand it: The Go and Kubernetes patent licenses terminate if you file litigation specifically regarding those projects-- if you sue anyone for violating some garbage collector patent in Go, you lose the patent licenses Go granted. Facebook's patent license terminates if you sue Facebook, subsidiaries, or "any party relating to the Software" for any patent infringement. It doesn't have to be related to React.
- diakritikal 9y ago
  Thanks, very interesting difference. Google's grant protects (or tries to protect) specific tech, Facebook's grant protects (or tries to protect) as an umbrella, the organisation. Well - can't comment on how effective it will be as a disincentive to sue, but I do feel at the same time if it does succeed in that goal it will almost certainly act equally as a negative factor in adoption; because it's my opinion that this is creating a soft-walled garden around Facebook open source tech.